A Twin gradient method for unconstrained optimization
Abstract
We propose a new strategy for gradient-based unconstrained optimization, involving two parallel sequences of iterates that cooperate to determine their stepsizes via a \textit{Twin-Step} principle. Rather than minimizing the objective function individually, the algorithm selects steplengths that minimize the Euclidean distance between the two gradient based processes occurring simultaneously at each iteration. The theoretical analysis shows that the convergence of the mutual distance is governed by the angle between the search directions. In particular the effectiveness of the overall process degrades as the directions approach parallelism. To ensure robustness against collinearity, we introduce a hybrid framework, Twin-ABBmin, which switches to the Adaptive Barzilai--Borwein method when the geometric cooperation becomes ineffective. Extensive and very promising numerical results evidence that the Twin phase creates favorable initial conditions for subsequent BB-type iterations.
Related
- A simple adaptive accelerated optimistic dual averaging method for unconstrained convex optimization in p-norm geometry
- Optimization method for flexible scheduling of digital twin industrial loads based on load side potential elasticity
- Global optimization by random perturbation of the gradient method with a fixed parameter
- A Stochastic Proximal-Gradient Method for Regularized Generally Constrained Continuous Optimization
- A Proximal Gradient Method for Nonconvex Nonsmooth Optimization with Applications
BibTeX
@misc{magistris2026twin,
title = {{A Twin gradient method for unconstrained optimization}},
author = {Anna De Magistris and Michiel E. Hochstenbach and Gerardo Toraldo},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2607.11575},
eprint = {2607.11575},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2607.11575},
}
RIS
TY - UNPB
TI - A Twin gradient method for unconstrained optimization
AU - Anna De Magistris
AU - Michiel E. Hochstenbach
AU - Gerardo Toraldo
PY - 2026
DA - 2026/10/30
PB - arXiv
DO - 10.48550/arxiv.2607.11575
UR - https://arxiv.org/abs/2607.11575
AB - We propose a new strategy for gradient-based unconstrained optimization, involving two parallel sequences of iterates that cooperate to determine their stepsizes via a \textit{Twin-Step} principle. Rather than minimizing the objective function individually, the algorithm selects steplengths that minimize the Euclidean distance between the two gradient based processes occurring simultaneously at each iteration. The theoretical analysis shows that the convergence of the mutual distance is governed by the angle between the search directions. In particular the effectiveness of the overall process degrades as the directions approach parallelism. To ensure robustness against collinearity, we introduce a hybrid framework, Twin-ABB_min, which switches to the Adaptive Barzilai--Borwein method when the geometric cooperation becomes ineffective. Extensive and very promising numerical results evidence that the Twin phase creates favorable initial conditions for subsequent BB-type iterations.
ER -